Psychophysiology-aided Perceptually Fluent Speech Analysis of Children Who Stutter

Area of Science:

• Speech-Language Pathology
• Biomedical Engineering
• Developmental Psychology

Background:

• Stuttering in young children can be influenced by situational physiological arousal affecting speech production.
• Identifying speech-motor-control factors is crucial for understanding stuttering disfluencies.
• Existing methods may not fully capture the dynamic interplay between physiology and speech in children.

Purpose of the Study:

• To present a novel approach, PASAD, for detecting changes in perceptually fluent speech acoustics in young children.
• To leverage real-time physiological responses to enhance speech signal analysis for identifying stuttering causes.
• To map distinct speech attributes in children who stutter (CWS) to their speech-motor-control factors.

Main Methods:

• Developed and applied the PASAD approach, utilizing a Hyper-Network structure to extract temporal speech importance.
• Collected simultaneous speech and physiological sensing data from 73 preschool-age children (CWS and CWNS).
• Analyzed speech-motor-control factors by mapping unique speech attributes identified by PASAD.

Main Results:

• PASAD effectively identifies speech attributes specific to CWS during fluent speech.
• The approach successfully maps these attributes to the speaker's speech-motor-control factors.
• Comprehensive evaluations demonstrate PASAD outperforms state-of-the-art multi-modal approaches.

Conclusions:

• PASAD offers a novel, adaptive, and robust method for analyzing speech in children, particularly those who stutter.
• The extracted knowledge enhances understanding of speech-motor control and stuttering development.
• PASAD's real-time execution capability makes it a valuable tool for clinical and research applications.